TheBloke
Published quantized GGUF, GPTQ and AWQ versions of open language models on Hugging Face
TheBloke is a public handle on Hugging Face, where the profile lists quantisation and fine tuning as interests and hosts quantized versions of open language models in the GGUF, GPTQ and AWQ formats. TheBloke's repository for Meta's Llama 2 7B, for example, offers GGUF files at quantization levels from Q2_K to Q8_0 and lists llama.cpp among the local clients and libraries that can run them. Earlier GGML repositories, such as the one for Koala 7B, say some of their files were quantized with an older version of llama.cpp, and in May 2023 TheBloke authored a merged commit to llama.cpp that let its convert.py script handle models stored in a single pytorch_model.bin file.1234
Work
llama.cppOpen-source code
Authored the May 16, 2023 commit "convert.py: Support models which are stored in a single pytorch_model.bin" (#1469). The repository is MIT-licensed.45
Quantized model repositories on Hugging Face (external site: huggingface.co)Open weights
Publishes GGUF, GPTQ and AWQ quantizations of open models under the TheBloke account. The Llama-2-7B-GGUF card says the repository contains GGUF files for Meta's Llama 2 7B under the llama2 license, and lists llama.cpp, LM Studio and other clients that support the format. The koala-7B-GGML card says its original quantisation methods were quantized using an older version of llama.cpp so that they stay compatible with llama.cpp as of commit 2d5db48.123
Earlier roles
- Contributor (convert.py, 2023), llama.cpp project4
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.